Roadmap

Reconstructed from the roadmap v0.2 conversation; the original document still needs exporting from claude.ai into this file (see TODO in decisions). Structure: phases with go/no-go gates. A NO-GO at any gate is a success of the method, not a failure of the project.

Phase 0 — Knowledge sprint (now — time-boxed, per DEC-007)

Before any marina hears from us, build enough knowledge to be an informed counterpart:

• 0.1 Course & licence. Pick and book the boating course (carta de Marinheiro as baseline; evaluate going straight to Patrão Local). Knowing the licence path also answers what we’re legally allowed to do (SCN-001’s daytime-return constraint).
• 0.2 Micro-rental. Rent a moored houseboat for 2–3 nights — the cheapest possible first data point for SCN-002/SCN-003/SCN-004 before the full January POC.
• 0.3 Market survey. Populate the boat archive: more models, real pricing, delivery conditions to Lisbon, and what the used market actually offers.
• SPK-1 — Email the marina(s) moves to the exit of Phase 0: still the question that can kill the project (SCN-012), now asked from an informed position. With DEC-004 open, ask both Parque das Nações and Oeiras.

Gate G0: course booked or done, nights slept aboard ≥ 2, archive has ≥ 5 models with prices — then SPK-1 goes out.

Phase 1 — Discovery: measure production

• 1.1 User research — 8–10 interviews with actual liveaboards in Lisbon-area marinas (interview script to write).
• 1.2 Telemetry of current life — 3–4 weeks on land: humidity/temperature per room (ESP32/M5Stack), smart plugs on key loads, activity log. This is the baseline every boat scenario gets compared against.
• 1.3 Requirements matrix — activity → requirement, MoSCoW priorities. The matrix constrains the boat, not the other way around.

Gate G1: matrix exists, is data-backed, and DEC-004 has an answer (or a justified deferral).

Phase 2 — Platform selection

• Shortlist from the boat archive filtered through the matrix and the scenario suite.
• Visits, sea trials where possible, real quotes (boat + berth + insurance + delivery).

Gate G2: no scenario still OPEN in any Must dimension; total cost model (SCN-013) complete.

Phase 3 — POC in staging (target: January)

• Rent/borrow a comparable boat and live aboard: weeks not days, normal workdays included, Marley aboard for at least a few nights, same sensors deployed for land-vs-boat comparison.
• Acceptance criteria agreed before stepping aboard: productivity, sleep, humidity, Marley.

Gate G3 — go/no-go. NO-GO in February costs ~€400. The same discovery in June costs ~€140k.

Phase 4 — Migration

• Purchase, survey, insurance, berth contract, move aboard.
• Rollback plan: the land home stays available through one full winter.

Phase 5 — Production

• Telemetry aboard, costs logged, scenarios re-run against reality, and the log tells the story.

Immediate backlog